VDS Browser Automation

VDS Browser Automation lets the VDS agent work in your own, already signed-in browser for a task you approve: open pages, read them, fill in filters and forms, click, scroll and take screenshots, without you ever giving VDS your passwords, cookies or session tokens. This extension works only together with a VDS account. It does nothing until you connect it to a VDS workspace. HOW IT WORKS 1. Install the extension, open the Browser Automation connector page in VDS and click "Connect this browser". Confirm the connection in the extension's window and in Chrome's permission prompt. 2. Ask the VDS agent for a task that needs your browser. At the start of every task a card in the VDS conversation asks you to start it and to choose how much you want to approve: every step (Manual), only steps that can change data (Auto-read), or let the AI decide which steps need you (Auto). 3. The agent works step by step. Steps that need your approval show a card in VDS with the action, the target and the text or script it will use. You can approve, reject or stop the task at any time. WHAT IT CAN DO - Open a web address in a new background tab, or work in tabs you already have open and prepared (filters, search results, expanded sections). Several tabs can be part of one task. - Read a page (its text, form values and links), take screenshots, scroll, hover and wait for content. - Click, type, select options, press keys, answer page dialogs and run short scripts in the page, when your VDS administrator has turned on interaction for the connector. By default the connector is read-only: the agent can only open and read pages and list your tabs. - List your open tabs (titles and addresses) so the agent can pick the right one. Sensitive values in addresses, such as tokens, session IDs and passwords, are replaced with "[redacted]" before they leave your browser. Incognito tabs and browser pages are never listed. BUILT-IN LIMITS - Never types into password fields. When a sign-in page appears, the agent asks you to sign in yourself. - The VDS conversation page that started the task can never be operated, so the agent cannot approve its own steps. - Closing one of your own tabs always asks you first, in every mode. - Chrome shows its "is debugging this browser" bar while a task runs. Clicking Cancel on that bar stops the task immediately. A task also ends when the answer is complete, when you click Stop in VDS, or after 10 minutes without activity. - Works only for VDS workspaces you connected. A page cannot talk to the extension unless it is one of them. WHAT IS SENT, AND WHAT IS NOT Sent only to the VDS workspace you connected, and only as the result of a step the agent was allowed to run: the text of the page, screenshots, results of scripts the agent ran, and tab titles and addresses. The extension itself never reads or sends your saved passwords or cookies, and never sends anything to anyone but the VDS workspace you connected. Scripts that the agent runs in a page you approved run with that page's own access, so the agent can see what that page can see. Nothing is sent when no task is running. YOU STAY IN CONTROL - Approval: you choose the approval mode for every task. On an approval card you can also choose "Approve and allow on this site" to let further changing steps on that site run without asking for the rest of the task. - Site access: Chrome asks you for access to websites when you connect. You can remove it at any time on chrome://extensions. - Connected workspaces: see and disconnect any VDS workspace in the extension options. - Step log: the options page lists the last 500 steps (time, workspace, tool, site without query string, result). Typed text, scripts and page content are never stored in the log, and you can clear it at any time. PERMISSIONS - debugger: the extension controls the tab of a task you approved through Chrome's debugging interface so it can read the page, click, type and take screenshots like a person would. Chrome shows its debugging bar while it is attached. - Access to websites (optional, requested only when you connect a VDS workspace): needed to connect the page of your VDS workspace to the extension. - Tabs: to list and open tabs of a task and to know when a tab you control closes or opens another tab. - Scripting: to register the small bridge script on your connected VDS workspace. - Storage: to keep your connected workspaces, the state of running tasks and the step log on your device. - Alarms: to end tasks that were left idle. VDS Browser Automation is published by VibeDecision. Privacy policy: https://vibedecision.com/privacy-policy